Abstract

Abstract – The Swedish Environmental Protection Agency is interested in obtaining current, detailed maps of mountain vegetation in Sweden. A project began in 2003 to map mountain vegetation using Landsat satellite data and field data from a new inventory called NILS (National Inventory of the Landscape in Sweden). NILS is the first country-wide systematic ground inventory to include the mountain area, and it is of interest to use these data for satellite data classification. Three Landsat scenes were topographically corrected and radiometrically corrected (using relative calibration from TERRA MODIS data). This allowed use of field data from all three scenes for classification. Supervised and unsupervised classification were tested and compared. Using NILS field data for accuracy assessment of the unsupervised classification, an overall accuracy for 19 vegetation classes was approximately 74%. The resulting supervised classification was better than expected given the limited amount of NILS field data. Using an increased number of NILS data, supervised classification may be a viable method for future classifications of mountain vegetation

Keywords

mountain vegetation; field inventory; Landsat

Published in

Publisher: Proceedings, ISRSE Symposium

Conference

31st International Symposium on Remote Sensing of Environment

SLU Authors

• Reese, Heather

  • Department of Forest Resource Management, Swedish University of Agricultural Sciences
    

• Nilsson, Mats

  • Department of Forest Resource Management, Swedish University of Agricultural Sciences
    

UKÄ Subject classification

Landscape Architecture
Environmental Sciences related to Agriculture and Land-use


Permanent link to this page (URI)

https://res.slu.se/id/publ/6208